Comparative Analysis of Nuclear Cross Sections in Monte Carlo Methods for Medical Physics Applications

Transcription

1 Comparative Analysis of Nuclear Cross Sections in Monte Carlo Methods for Medical Physics Applications Christopher T. Myers 1 Georgia Institute of Technology Bernadette L. Kirk 2 Luiz C. Leal 2 Oak Ridge National Laboratory

2 Overview Monte Carlo (MC) simulations are increasingly being implemented for medical physics applications. Radiotherapy 8 Medical Accelerator Modeling 9 Nuclear Medicine Imaging 10 MC offers a simple and controlled method to determine a quantity of interest.

3 Overview There are three sources of discrepancy in Monte Carlo simulations: Computer Method Data Purpose: Eliminate the discrepancy arising from differences in data used by EGSnrc and MCNPX.

4 Photon Cross Section Data MCNPX 1 Mcplib04 Incoherent scattering Coherent scattering Photoelectric Pair Production (includes triplet data) Coherent Form Factors Incoherent Scattering Functions Edge Energies Relative probabilities of shell ejection Yields Fluorescent energies Heating numbers

5 EL03 Electron Cross Section Data MCNPX Bremsstrahlung Radiative Stopping Powers Binding Energy and Shell Occupation (used for density effect corrections) Electron induced relaxation thresholds Auger e - emission energy Scattering information (angles, functions)

6 PEGS EGSnrc 2 Cross Section Data PEGS Uses XCOM evaluation to create data library Pegs4pepr.dat Photoelectric absorption K-edge Energy Pair Production Coherent Scattering Pgs4form.dat Coherent form factors Aprime.dat Empirical Bremsstrahlung correction for 14 elements 3

7 EGSnrc Cross Section Data EPDL 3 and XCOM 4 libraries are available for linear interpolation: Photoelectric Pair/Triplet production Rayleigh (coherent) scattering Still must use PEGS input file *EPDL is the same evaluation used for MCNPX ** Pair/Triplet taken from EPDL97 and not MCNPX

8 EGSnrc Cross Section Data The following libraries are also used when running EGSnrc incoh.data Contains shell occupations, binding energies, and Compton profile parameters Used to correct default Klein-Nishina Nishina data for Doppler broadening and binding effects photo-relax.data Contains shell binding energies. (Uses same method as MCNPX)

9 EGSnrc Cross Section Data spinms.data Contains ratios of e-e and e+ multiple elastic scattering distributions with spin taken into account msnew.data Contains multiple scattering angles eii_ik.data Electron impact ionization cross sections for each subshell with BE > 1 kev nist_brems.data Bremsstrahlung cross sections based on NIST 6,7 evaluations. photo_cs.data Formula fits (accurate w/in 1-2%) 1 2

10 Converting MCNP Data to EGS Libraries Use NIST evaluation for Bremsstrahlung data. Replace binding energies in photo_relax.data and incoh.data with values from ENDF/B-VI.8. Interpolate electron impact ionization data found in ENDF/B-VI.8 to the energy grid used in EGSnrc (BE 10 GeV)

11 Converting MCNP Data to EGS Libraries O K-Shell Ionization Cross Section Pb K-shell Ionization Cross Section Ln(sigma) ln(sigma) ln(e) Pb L1-Shell Ionization Cross Section Ln(E) ln(sigma) ln(e)

12 Converting MCNP Data to EGS Libraries 1. Replace EPDL cross sections from mcplib04. (to match energy grid) 2. Edit data used to create PEGS library. Create new.dat. file Photoelectric absorption cross sections K-edge energies for 1 Z 100. Pair production cross sections Coherent Scattering Import coherent form factors from mcplib04 into pgs4form.dat

13 EGS XS-file summary Converted PEGS4 Pegs4form.dat Pegs_xcom-full.dat photo_relax.data eii_ik.data photo_cs.data XCOM/EPDL Rayleigh Pair Production Photoelectric Unchanged PEGS4 aprime.dat msnew.data spinms.data nist_brems.data* incoh.data XCOM/EPDL triplet *Equivalent to MCNP data

14 Comparing Dose Distributions and Particle Creation Events EGSnrc dose compared to MCNPX dose in cylindrical regions from an isotropic, mono-energetic cylindrical source. Particle histories and run times for EGSnrc and MCNPX simulations are compared.

15 Radial Depth Dose 500 kev Photons 1.01 Ratio to MCNPX cm MOD EGS XS ORIG EGS XS

16 Depth Dose 1 MeV Photons 1.02 Ratio to MCNPX cm MOD EGS XS ORIG EGS XS

17 Depth Dose for Incident 5 MeV Photons Ratio to MCNPX cm Mod EGS XS ORIG EGS XS

18 Depth Dose for Incident 10 MeV Electrons EGSnrc vs MCNPX 1.01 Ratio to MCNPX cm MOD EGS XS ORIG EGS XS

19 Particle Histories for 1 MeV Photons 1-MeV Photon MCNPX EGSnrc EGS XS MCNPX XS Electrons from Source pair production Compton recoil photoelectric auger photon 0 x X auger electron 0 x X knock on/ Möller Bhabha x 0 0 Photons from Source Bremmstrahlung p-annihilation electron x-rays fluorescence TOTALS Photons Electrons CPU time 16.2 s 19 s 19 s

20 Particle Histories for 10 MeV Electrons 10 MeV Electron MCNPX EGSnrc EGS XS MCNPX XS Electrons from Source pair production Compton recoil photoelectric auger photon 0 X X auger electron 0 X X knock on/ Möller Bhabha x 0 Photons from Source Bremmstrahlung p-annihilation electron x-rays fluorescence TOTAL Photons Electrons CPU time 543 s 44.5 s 44.6 s

21 Particle History Comparisons The Knock-on/ on/möller scattering shows the largest discrepancy. (2-3 3 O.M.) Compton (Incoherent) scattering shows large differences could be due cross sections. MCNPX requires slightly less CPU time for incident photons. EGSnrc MUCH less CPU time for incident electrons.

22 Conclusions There are discrepancies in the nuclear data between EGSnrc and MCNPX that contribute to discrepancies in results obtained using higher energy particles. Ignoring entrance and exit regions, the data from MCNPX in EGSnrc shows a slightly better agreement with MCNPX than does EGSnrc data. Run time differences observed between EGSnrc and MCNPX appear to be due to large differences in the numbers of electrons created and transported.

23 Future Work Convert Incoherent scattering cross sections and determine impact on results and particle creation. Complete program that creates new EGSnrc data based on mcplib04, el03, EPDL97, and ENDF/B- VI.8 libraries. Perform sensitivity calculations on the data.

24 References 1. D. B. Pelowitz,, ed., MCNPX User's Manual, Version 2.5.0, LA-CP CP (April 2005). 2. Kawrakow, I and D.W.O. Rogers. The EGSnrc Code System:Monte Carlo Simulation of Electron and Photon Transport. National Research Council of Canada. (Sept. 21, 2006). 3. D. E. Cullen et al., EPDL97: the Evaluated Photon Data Library, 97 Version, UCRL-50400, Vol. 6, Rev. 5, The University of California, Lawrence Livermore National Laboratory, Livermore, CA (1997). 4. XCOM: Photon Cross Sections Database, National Institute of Standards and Technology (1998). 5. Cross Section Evaluation Working Group, ENDF/B-VI Summary Documentation (ENDF-201), BNL-NCS-17541, 8th Edition, National Nuclear Data Center, Brookhaven National Laboratory (2000). 6. S. M. Seltzer and M. J. Berger, Bremmstrahlung Spectra from Electron Interactions with Screened Atomic Nuclei and Orbital Electrons Nucl. Inst. Meth. Phys. B. 12, (1985). 7. S. M. Seltzer and M. J. Berger, Bremmstrahlung Energy Spectra from Electrons with Kinetic Energy from 1 kev to 10 GeV Incident on Screened Nuclei and Orbital Electrons of Neutral Atoms with Z = 1-100, Atomic Data and Nuclear Data Tables, 35, (1986). 8. T. R. Mackie, Applications of the Monte Carlo Method in Radiotherapy, pp in Dosimetry of Ionizing Radiation, Vol. III, Academic Press, New York (1990). 9. P. Andreo, Monte Carlo Techniques in Medical Radiation Physics, Phys. Med. Biol. 36, (1991). 10. H. Zaida, Relevance of Accurate Monte Carlo Modeling in Nuclear Medicine Imaging, Med. Phys. 26, (1999).